Savr.

Savr is a zero-waste, budget-smart, AI-native weekly meal planner built on a different premise than the incumbents (Mealime, PlateJoy): rather than recipe-first then shopping list, Savr starts from your pantry and your budget and works backward — generating a weekly plan that maximizes ingredient reuse, respects dietary needs, hits cuisine preferences, and stays under the weekly grocery budget.

The planning engine combines four AI capabilities into a single optimization pass: (1) Dietary-needs matching (recipes must fit restrictions first), (2) Ingredient reuse (according to user preference: maximize savings / balanced / maximize variety), (3) Cost estimation with budget enforcement, (4) Budget recovery — if a plan goes over, swap recipes to reduce cost. The system holds three constraints simultaneously: budget-first, zero-waste, time-aware.

Powered by GPT-4o-mini — chosen specifically for Structured Outputs support, speed, affordability, and reliability at generating consistent JSON from a template. The task profile (parse user constraints, select recipes, perform structured generation, return JSON) fits this model's strengths exactly without paying for frontier-tier reasoning. Automated test suite runs schedule-coverage tests (12: under-fill, hybrid retry recovery, retry exhaustion, outside-window retries, over-fill trim, exact-match) and invariant tests (17) on every change.

Targeting the AI-driven meal planning category projected to grow at 28.10% CAGR (2025–2034), driven by consumer demand for personalized health & wellness, customized meal plans, and informed food choices. B2C — primary persona: busy students, working professionals, and eco-conscious consumers with specific dietary restrictions who hate wasting food and want planning to feel lightweight and guided, not like another spreadsheet. Launch approach: limited friendly beta with colleagues, friends, and a small test group to validate usability before public release.